At 2:02 PM -0600 11/10/00, Eric Scharin wrote:
>This is starting to seem relevant to the thread of a few weeks back
>regarding the difference between statistical & practical significance.
>It may be that the 96 Palm Beach bad ballot numbers were statistically
>significant, but not practically so (wouldn't have had an impact on the
>outcome either way). The 00 results are definitely practically
>significant (otherwise there wouldn't be this debate), and could well
>be the difference between electing Bush or Gore.
>
>Again, I heard that the real difference in disqualified votes between
>96 & 00 is around 15,000 vs. 30,000, when you compare apples to apples.
>
>And I'd also like to enter a contrarian opinion on the incompetence of
>the voters. Incompetence or ignorance (even if that is the root cause
>of the misvotes) is currently not a disqualifier for involvement in the
>democratic process. If that were the case, they wouldn't provide
>assistance for illiterate voters (which they do) and (IMHO) they
>wouldn't allow 90% of the candidates to run for office.
>
>The discussions I've heard during the media coverage of this all have a
>disconcertingly political tinge to them. There seems to be a lack of
>debate based on principle. The principle I'm referring to the right of
>every eligible citizen to have their opinion heard and choice recorded.
>If the voting system in place in Palm Beach hampered this fair process,
>then it needs to be investigated in an even-handed way, considering all
>of the data available.
>
> - Eric
>
>-----Original Message-----
>From: [EMAIL PROTECTED]
>[mailto:[EMAIL PROTECTED]]On Behalf Of Reg Jordan
>Sent: Friday, November 10, 2000 12:25 PM
>To: [EMAIL PROTECTED]; dennis roberts
>Subject: Re: Stats on Palm Beach votes
>
>
>These 19,000 "spoiled" ballots account for about 0.3% of the vote count.
>Doesn't seem too high to me.
That's .3% of the vote count _in the entire state of Florida_, which
isn't the relevant statistic. The 19,000 "overvoted" ballots under
discussion were all in Palm Beach County, where the total vote was about
400,000---so this is 5% of the total, and about 8% when you add in the
10,000 "undervoted" ballots. Tossing out 1 in 12 ballots is pretty high
in my book.
The problem is partly one of technology. Travis County (TX) where I live
used to have ballots of this kind; I always thought them a pain,
difficult to use, and I used to religiously check my ballot after I
withdrew it from the "book" to see that all the "chad"--little bits of
cardboard that get punched out when you vote a position--were completely
removed, and also that the numbers printed on the ballot agreed with the
numbers in the book. I bet not one in a hundred voters did as I did, and
it certainly shouldn't be required, but this is very poor technology and
it is highly prone to error. And as the "butterfly" ballot shows, there
is also a very big human interface problem with this technology.
Massachusetts has banned this technology statewide because of its
unreliability (after a contested election, BTW).
Here's something I posted elsewhere that is relevant:
----
>I heard a strange statement on NPR this morning. Someone said
>that the problems with the ballots were not all that unexpected,
>15,000 had to tossed in Florida in 96.
There's been some confusion on this point. The number in 96 was 14,000,
of which over half were "undervotes", that is, ballots that had no vote
for any candidate in some position (it's not clear if this is all
positions or only Presidential ones). Less than half, under 7,000, were
"overvotes", where two candidates in the same position were voted for.
The 19,000 ballots mentioned in connection with this election are all
"overvotes", and are all for President. Thus, you are comparing under
7,000 with 19,000, not 14,000 with 19,000. 19,000 is about 5% of the
vote in that county, which I think unacceptably large. 7,000 is still
pretty big, though. It appparently didn't make a difference to the
result in 1996, which is probably why nothing was done about it. This
time the difference is crucial, so people are paying attention.
There are also 10,000 "undervotes" in this election. Given the punch
card technology being used, it is likely that a significant number of
these were votes where the punch did not go completely through and the
optical reader failed to read the vote. (Indeed, one of the reasons that
the recount changes is that sometimes the partly punched-through "chad"
falls off the second time through and the vote is successfully counted).
The purpose of the proposed hand count is to see if there are instances
where this has happened by physically inspecting each ballot. Bush's
folks don't like this because the areas of Florida where the punch card
technology is used are heavily Democratic. The Northern part of the
state uses a mark-sense technology similar to that used for SAT tests,
which is not prone to this sort of error.
Another kind of error also produces undervotes, and that is when the
voter puts the ballot in the machine upside down; such a situation
produces no punches at all, but it does produce dimples on the ballot
when the voter attempts to punch it, which can be read. Election
officials consider this an easy problem to remedy, in fact.
The 19,000 overvotes probably can't be rescued by any reasonable method.
One hopes that Florida will abandon the punch card technology and go to
a more reliable one like the mark-sense technology used in Northern
Florida.
This information courtesy of Salon.com, which published it today
(Saturday).
Bill
--
Bill Jefferys/Department of Astronomy/University of Texas/Austin, TX 78712
Email: replace 'warthog' with 'clyde' | Homepage: quasar.as.utexas.edu
I report spammers to [EMAIL PROTECTED]
Finger for PGP Key: F7 11 FB 82 C6 21 D8 95 2E BD F7 6E 99 89 E1 82
Unlawful to use this email address for unsolicited ads: USC Title 47 Sec 227
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
